Abstract
OBJECTIVE: This study aimed to examine whether the 10-item Risk Score List (RSL) accurately predicts delirium in patients admitted to inpatient hospice care and whether this instrument can be simplified. Determining the risk for developing delirium can help to treat these patients in a timely manner. METHODS: This was a retrospective medical record study in patients who died in 2019 or 2020 in three hospices. Predictive values were examined using Cox regression analysis, crosstabs, and C-statistic. RESULTS: In total, 240 patients were included. Median age at admission was 78 (IQR 70-84) years. Primary diagnosis most often was cancer (n = 186, 78%); 173 (72%) patients had an increased risk of delirium according to RSL, of whom 120 (69%) developed delirium. Overall, 147 (61%) patients developed delirium. The RSL significantly predicted future delirium (HR 3.25, CI 1.87-5.65, p < 0.01) and had a sensitivity of 85%, a specificity of 43%, positive predictive value of 62%, negative predictive value of 73%, and a C-statistic of 0.64. Simplifying the RSL to four items still significantly predicted future delirium, with similar predictive values. CONCLUSION: Delirium occurs in more than half of patients admitted to hospice care. The RSL can be simplified to four items, without compromising on predictive accuracy.